Using consistency-driven pairwise comparisons in knowledge-based systems

Abstract

Knowledge acquisition often requires the assessment of qualitative stimuli (e.g., criteria or priorities) such as public safety or the degree of environmental damage. Weights, reflecting the relative importance of the objectives concerned are one of the most commonly used solutions for this kind of data. Subjective judgments involve inaccuracy (which is difficult to control) and consistency of judgments (which can be measured and used to influence the accuracy). The consistency-driven pairwise comparisons method is based on an inconsistency index and its use as a validation technique. Knowledge acquisition using the consistency-driven pairwise comparisons method, when supported by properly designed software, contributes to the improvement of the quality of knowledge-based systems.